Struct Corpus 

pub struct Corpus {
    pub name: String,
    pub root_path: PathBuf,
    pub images: Vec<CorpusImage>,
    pub metadata: CorpusMetadata,
}

A corpus of test images.

Fields

name: String

Name of the corpus.

root_path: PathBuf

Root path of the corpus.

images: Vec<CorpusImage>

Images in the corpus.

metadata: CorpusMetadata

Metadata about the corpus.

Implementations

impl Corpus

pub fn new(name: impl Into<String>, root_path: impl Into<PathBuf>) -> Self

Create a new empty corpus.

pub fn discover(path: impl AsRef<Path>) -> Result<Self>

Discover images in a directory.

Recursively scans the directory for supported image formats (PNG, JPEG, WebP, AVIF).

pub fn get_dataset(dataset: &str) -> Result<Self>

Get corpus dataset, downloading if necessary.

When the corpus feature is enabled (default), uses the codec-corpus crate for automatic download and caching. Otherwise, checks if the path exists locally.

Arguments

• path - Dataset path (e.g., “kodak”, “clic2025/training”)

Example

// With corpus feature (default): downloads and caches automatically
let corpus = Corpus::get_dataset("kodak")?;

// Discovers images in the cached directory
println!("Found {} images", corpus.len());

pub fn discover_or_download( path: impl AsRef<Path>, _url: Option<&str>, _subsets: Option<&[&str]>, ) -> Result<Self>

Discover or download a corpus on demand (legacy, requires local path).

If the path exists, discovers images. Otherwise returns an error. Use get_dataset() when the corpus feature is enabled for automatic downloads.

Arguments

• path - Local path for the corpus
• _url - Ignored (for backward compatibility)
• _subsets - Ignored (for backward compatibility)

pub fn download_dataset(dataset: &str) -> Result<Self>

Download a specific dataset (replaces download_subset).

With the corpus feature enabled, this uses codec-corpus for caching.

Example

let corpus = Corpus::download_dataset("kodak")?;

pub fn get_or_download(preferred_path: impl AsRef<Path>) -> Result<Self>

Get corpus from local paths (legacy method).

Checks common locations for existing corpus:

1. The specified path
2. ./codec-corpus
3. ../codec-corpus
4. ../codec-comparison/codec-corpus

When the corpus feature is enabled, use get_dataset() instead for automatic download and caching.

pub fn load(path: impl AsRef<Path>) -> Result<Self>

Load a corpus from a JSON manifest file.

pub fn save(&self, path: impl AsRef<Path>) -> Result<()>

Save the corpus to a JSON manifest file.

pub fn len(&self) -> usize

Get the number of images in the corpus.

pub fn is_empty(&self) -> bool

Check if the corpus is empty.

pub fn filter_category(&self, category: ImageCategory) -> Vec<&CorpusImage>

Filter images by category.

pub fn filter_format(&self, format: &str) -> Vec<&CorpusImage>

Filter images by format.

pub fn filter_min_size( &self, min_width: u32, min_height: u32, ) -> Vec<&CorpusImage>

Filter images by minimum dimensions.

pub fn split(&self, train_ratio: f64) -> (Vec<&CorpusImage>, Vec<&CorpusImage>)

Split the corpus into training and validation sets.

Uses a deterministic split based on checksum to ensure reproducibility.

Arguments

• train_ratio - Fraction of images to include in training set (0.0-1.0).

pub fn compute_checksums(&mut self) -> Result<usize>

Compute checksums for all images that don’t have them.

pub fn find_duplicates(&self) -> Vec<Vec<&CorpusImage>>

Find duplicate images by checksum.

pub fn update_category_counts(&mut self)

Update category counts in metadata.

pub fn stats(&self) -> CorpusStats

Get statistics about the corpus.

Trait Implementations

impl Clone for Corpus

fn clone(&self) -> Corpus

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Corpus

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for Corpus

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for Corpus

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.